Depleted mantle-plume geochemical signatures: No paradox for plume theories

Depleted mantle-plume geochemical signatures: No paradox for plume theories Andrew C. Kerr Andrew D. Saunders Department of Geology, University of Leicester, University Road, Leicester LE1 7RH, United Kingdom John Tarney Neil H. Berry Victoria L. Hards Department of Geological Sciences, University of Durham, South Road, Durham DH1 3LE, United Kingdom

ABSTRACT High-MgO liquids erupted in ocean-island settings and in some continental flood-basalt provinces commonly preserve a ‘‘depleted’’ composition, in terms of both highly incompatible trace elements and isotope ratios. These observations strongly imply that their source is also compositionally depleted. However, in at least one case (Iceland and the North Atlantic volcanic province), it can be shown that this depleted source is not the same as that feeding the present-day North Atlantic mid-ocean ridge. The depleted source must also have been much hotter than the mid-ocean ridge basalt (MORB) source to account for the volume of melt and primitive composition of some magmas that were generated. This depleted character, then, is an intrinsic component of mantle plumes, originating from the deep mantle. We propose that mantle plumes consist of a mixture of enriched or fusible streaks in a depleted, refractory matrix; preferential extraction of the enriched component occurs close to the plume axis. The depleted residue from this melting remains in the upper mantle and may therefore be a major contributor to the source region of MORB. INTRODUCTION It is widely accepted in the geochemical literature that thermally buoyant mantle plumes which rise from a boundary layer within the mantle (e.g., White and McKenzie, 1989; Campbell and Griffiths, 1990) have trace element contents and Sr-Nd isotope signatures that are enriched relative to the source region of mid-ocean ridge basalt (MORB). It is observed that the decompression melt products of these plumes are commonly more enriched in incompatible trace elements and have Sr-Nd and Pb isotope ratios distinct from those of MORB. As a result, several chemically distinctive, enriched1 plume components have been postulated in order to explain the geochemical signatures of intraplate oceanic basalts (e.g., White, 1985; Zindler and Hart, 1986; Weaver, 1991). One additional component, depleted MORB mantle, is generally regarded to be the source of MORB. In this paper we challenge the assumption that all plumes are enriched (cf. Anderson, 1994), and we present evidence that strongly supports the existence of plumes having substantial depleted component(s). These melt to produce magmas with compositional similarities to normal (N-type) MORB. We review the possible origins of this depleted end member and probable mechanisms for its incorporation into an upwelling plume. Anderson (1994) suggested that the existence of depleted mantle-plume signatures ‘‘is a major paradox for plume theories’’; however, we show that this is not the case and that depleted geochemical signatures of plumes are entirely explicable within the framework of conventional plume models. 1 In the remainder of the paper, the terms ‘‘enriched’’ and ‘‘depleted’’ are relative to bulk earth, in terms of trace element contents and radiogenic isotopes.

Geology; September 1995; v. 23; no. 9; p. 843– 846; 2 figures.

DEPLETED MANTLE IN OCEANIC PLUME SOURCES The Icelandic mantle plume has been producing basaltic melts for the past 63 m.y. (White 1988) and possibly longer (Lawver and Mu ¨ller, 1994). Today this plume impinges on the base of the lithosphere below the Mid-Atlantic Ridge. Two main magma types are found on Iceland: on-axis tholeiites and picrites that erupt above relatively thin lithosphere, and off-axis alkalic basalts that form below thicker lithosphere (He´mond et al., 1993; Hards et al., 1995). The tholeiites and picrites are generally much more depleted (LaN/ SmN , 1 and εNd . 17.1) than the alkalic basalts (LaN/SmN . 1 and εNd , 17.8; He´mond et al., 1993). Some authors attribute this depletion to the melting of MORB-source asthenosphere that has been entrained into the plume (Schilling et al., 1982; Elliott et al., 1991; Hart et al., 1992). More recently, He´mond et al. (1993) and Hards et al. (1995) showed that there are subtle elemental and isotopic differences between Icelandic volcanic rocks and North Atlantic MORB. Pb isotopes, particularly 207Pb/204Pb and 208Pb/204Pb ratios (Fig. 1) clearly show these compositional differences (Thirlwall et al., 1994). Although Figure 1 supports plume-MORB mixing along the Reykjanes Ridge (Schilling, 1973), it also precludes the involvement of any significant volume of local depleted N-type MORB mantle in the source region of Icelandic volcanic rocks, because Icelandic basalts and North Atlantic MORB define two, subparallel arrays. Contamination of N-type MORB basalts with altered Icelandic crust cannot explain this difference in Pb isotope systematics simply because Icelandic crust is too young (Hards et al., 1995). Therefore, it would seem that rather than entrained MORBsource mantle, the depleted source region of the Icelandic picrites and olivine tholeiites is an integral part of a heterogeneous plume (Hards et al., 1995).

Figure 1. Plot of 207Pb/204Pb vs. 208Pb/204Pb, showing fields for Icelandic volcanic rocks and North Atlantic normal (Ntype) MORB away from influence of transform faults and known mantle plumes. Sources for Reykjanes Ridge data— Sun et al. (1975) and references in Hards et al. (1995). 843

The Late Cretaceous komatiites and picrites of Gorgona Island, Colombia, have εNd . 10 and LaN/SmN ,, 1 and are derived from primary magmas with .18% MgO; there is little evidence to suggest that the mantle source region was ‘‘wet’’ (Kerr et al., 1995). Therefore, Kerr et al. (1995) argued that the depleted source of the Gorgona ultramafic lavas was much too hot to represent ambient upper (i.e., MORB source) mantle. They have proposed that the plume responsible for the magmatism on Gorgona was itself markedly heterogeneous, containing both depleted and more enriched components. Many of the lavas erupted on Hawaii have enriched trace element signatures and radiogenic isotopes. Nevertheless, various authors have noted that despite their origin in a hot mantle plume, a significant number of Hawaiian basalts have been derived from a relatively depleted source region (Budahn and Schmitt, 1985; Stille et al., 1986). It has been proposed that this depletion could be due to interaction of plume melts with depleted MORB-like lithosphere (e.g., Stille et al., 1986). CONTINENTAL FLOOD-BASALT PROVINCES Although many lavas erupted in continental settings are contaminated by lithospheric material, most flood-basalt provinces appear to have a few relatively uncontaminated lavas, which enable us to get some idea of the composition of the plume end member. MacDougall (1988) has noted that the radiogenic isotope signatures of many continental flood-basalt (CFB) provinces show that at least one component in their source region has a long-term history of depletion. The 250 Ma lavas of the Siberian Traps represent one of the largest outpourings of continental flood basalt during the Phanerozoic. Picrites with depleted signatures (εNd 5 14 to 17.3 and LaN/SmN , 1.3; Lightfoot et al., 1993) are found near the base of the succession. In contrast, the lowest εNd so far reported from the basalts is 12 (Lightfoot et al., 1993). These observations strongly suggest that the picrites were derived from within the hot upwelling plume and that this source contained a depleted component. Within the early Tertiary Deccan CFB province, the lavas of the Ambenali Formation are generally regarded as being the least contaminated by lithospheric crust or mantle (maximum εNd of 1 7.1; Mahoney, 1988). However, Mahoney (1988) showed that the Ambenali basalts, despite their depleted Nd isotope signatures, have 208 Pb/204Pb and 207Pb/204Pb ratios that rule out a melt input from both Indian Ocean MORB-source mantle and the present-day Reunion plume. The Early Jurassic Karoo, South Africa, CFB province also contains examples of relatively depleted, plume-related, lavas and dikes. MORB-like dolerite dikes have been found at Rooi Rand, South Africa (εNd up to 14; Duncan et al., 1990). Ellam and Cox (1991) proposed that the enriched picritic basalts from Nuanetsi were mixtures of enriched small melt fractions from the lithospheric mantle and a plume component with a depleted MORB-like composition (εNd 5 110). The early Tertiary North Atlantic province is believed to result from decompression melting associated with the initial arrival of the head of the Icelandic plume below the lithosphere at 63 Ma (White, 1988). Kerr (1995) showed that basalts from throughout the province, which are uncontaminated with continental lithosphere, have a depleted signature and appear to have been derived from a depleted source region within the plume. The picrites of West Greenland and Baffin Bay represent some of the most primitive lavas found in the North Atlantic Tertiary province. They are also some of the most depleted, with MORB-like light-REE patterns and εNd 5 17 to 19 (O’Nions and Clarke, 1972; 844

Holm et al., 1993). Holm et al. (1993) proposed that the enriched plume component in these lavas had been diluted by MORB-source mantle, during the ascent of the plume. Depleted picrites are commonly found at the base of volcanic successions. This can be explained by poorly developed magma chambers in the early stages of volcanism; from such magma chambers, high-MgO melts pass relatively quickly through the crust to be erupted as picrites. As the volcanic system evolves, development of more magma chambers provides a greater chance for picrites to fractionate and, in the continental realm, become contaminated with (enriched) lithospheric material. We have attempted to show above that lavas from both currently active plumes and from plume-related CFB provinces display evidence for depleted components in their source region(s). Moreover, it is the most-magnesian lavas within these provinces that display the greatest degree of depletion. This is true of Iceland, Gorgona, the Siberian Traps, and the North Atlantic Tertiary province. It is also significant that Archean komatiites, arguably the most magnesian lavas erupted in Earth’s history, are also relatively depleted, having positive εNd values and LaN/SmN ,, 1 (Campbell and Griffiths, 1992). SOURCE OF THE DEPLETED COMPONENT AND ITS INCORPORATION INTO PLUMES There is a growing acceptance in both the geochemical and geophysical literature that mantle plumes must rise from a boundary layer within Earth’s mantle, either the 670 km discontinuity or from just above the core-mantle boundary (D0). Relatively high levels of primordial 3He, in comparison to MORB, and the inferred mantle potential temperatures (up to 200 8C hotter than ambient upper mantle) suggest that at least some of the most vigorous present-day plumes (e.g., Iceland and Hawaii) must have a hot, deep, mantle source region (Davies and Richards, 1992). There are several possible models to explain where plumes acquire these depleted components: 1. Entrainment of MORB-source upper mantle. Although, as noted above, Pb isotopic constraints rule out this possibility for Iceland and the Deccan CFB province, it must be taken into account in any discussion of depleted-plume signatures. Griffiths and Campbell (1990), Hart et al. (1992), and Hauri et al. (1994) have shown that when starting plumes and steady-state plumes rise from the core-mantle boundary, the material entrained into the ascending plume head will be predominantly from the lower rather than the upper mantle. Richards and Griffiths (1989) have shown that significant entrainment of upper-mantle material can occur when the plume conduit is deflected horizontally by larger than usual shear flow velocities in the upper mantle. It is also important to remember that smaller amounts of melt will be generated in the entrained mantle, relative to the rising plume. Even though the hot plume has the potential to heat up significant volumes of MORB-source mantle around its margins, this plume-heated upper mantle will not be as hot as the plume itself and so, at comparable pressures, will undergo only small-scale decompression melting. Although a possible explanation, this model does not explain why the hottest magmas erupted during plume-related volcanism are also the most depleted and why they sometimes have high 3He/4He ratios. On the basis of evidence from Sr-Nd isotopes and rare earth elements, Anderson (1994) proposed that depleted picrites and komatiites come from the same depleted reservoir as MORB. He further suggested that melts from this depleted MORB source could be contaminated with ‘‘shallow-level’’ enriched mantle to produce enriched plume signatures. However, both high 3He signatures and the excess mantle heat required to produce high-MgO (.15 wt%) GEOLOGY, September 1995

Figure 2. Schematic representation of plume involvement in mantle processes. Residual subducted slab accumulates in lower-mantle D( layer (A) and at 670 km discontinuity (B). C—Ascending plumes entrain surrounding mantle. D—Plumes impinge beneath thick oceanic lithosphere, generating ocean-island basalt from small-degree-ofmelting mantle melts, e.g., St. Helena. E—Passive upwelling of convecting upper mantle generates N-type MORB. F—Plume ascending beneath ridge forms oceanic plateau, including some depleted high-degree-of-melting picritic melts (e.g., Iceland). G—Plume arising beneath rifted ‘‘thin-spot’’ continental lithosphere generates voluminous continental flood basalts with some depleted picritic melts (e.g., West Greenland).

magmas pose fundamental problems for Anderson’s model. Hart et al. (1992) also showed that ocean-island basalts have Sr-Nd-He206 Pb isotopic systematics that, on the whole, do not require a significant input from MORB-source asthenosphere. Additionally, as noted above, the 207Pb-208Pb systematics of Icelandic (Fig. 1) and Deccan basalts (Mahoney, 1988) rule out any contribution to their melts from MORB-source upper mantle. 2. Entrainment of depleted material from the lower mantle. We have already noted that the heat and the chemistry of most depleted, plume-related lavas require a deeper source than the upper mantle. Griffiths and Campbell (1990) and Hauri et al. (1994) suggested that plumes rising from the core-mantle boundary have the potential to entrain lower-mantle material. The lower mantle thus represents a possible reservoir of this depleted mantle. Recent studies of diamond inclusions also lend support to the idea that the lower mantle contains substantial depleted components (Kesson and Fitzgerald, 1991). 3. Depletion in the mantle source. The plume source region could itself contain a depleted component as well as an enriched component. This source region could be the 670 km discontinuity, but for larger, more powerful plumes with high 3He/4He ratios, it is more likely that the source region is D0. In the latter case D0 can provide both the necessary heat and possibly 3He from degassing of the core. 3He can also be derived from primitive undegassed lower mantle. DISCUSSION Thus far we have sought to show only that depleted plume signatures exist and that this depletion is not primarily caused by the incorporation of MORB-source mantle into the plume. We now turn to the ultimate origin of the depleted plume component(s). It has been proposed by several authors (Hofmann and White, 1982; White, 1985; Zindler and Hart, 1986; Weaver, 1991) that the source region of mantle plumes (probably at the D0 layer) is fed by subducted oceanic crust, with variable amounts of subducted abyssal or continental sedimentary material to account for the EM1 and EM2 signatures. The lower oceanic lithosphere consists of cumulate rocks and residues from melting which would be more trace element depleted than the oceanic crust (Saunders et al., 1988). It has been suggested that this oceanic lithosphere separates from the overlying crust as the slab passes through the mantle (Hofmann and White, 1982; Weaver, 1991). However, if more depleted, refractory oceanic lithosphere is also transported down to the lower mantle, then this process would provide a possible explanation of how plumes obtain their depleted signatures. Recent work suggests that some slabs do not pass through the 670 km discontinuity, but instead are deflected and flatten along it GEOLOGY, September 1995

(Phipps-Morgan and Shearer, 1993; Tackley et al., 1993). Tackley et al. (1993) proposed that this slab material, which builds up at 670 km depths, eventually becomes gravitationally unstable and ‘‘avalanches’’ into the lower mantle, to accumulate at D0. This mechanism also allows depleted oceanic lithospheric mantle to accumulate with the oceanic crust at the 670 km discontinuity and so also be a part of the avalanche to the lower mantle and D0 (Fig. 2). Thus, all plumes have the potential to contain at least two components; a fertile component consisting of oceanic crust and its associated sediments and a depleted component possibly derived from oceanic lithospheric mantle. In the oceanic realm, many weaker plumes, and those that do not come up near a ridge, have enriched signatures (e.g., Cape Verde, Reunion, and St. Helena). In contrast, those plumes that impinge on the base of the lithosphere near a ridge (e.g., Iceland), or a lithospheric thin spot (e.g., West Greenland), result in more extensive partial melting. These higherdegree-of-melting liquids are more depleted and more MgO-rich than ocean-island basalt erupted in an intraplate setting (Fig. 2). It is possible that the enriched components are dispersed in plumes as more fusible streaks or blobs set in a more depleted, refractory matrix (cf. Zindler et al., 1984; Fitton and James, 1986). Where the base of normal oceanic lithosphere restricts the degree of melting by placing a ‘‘lid’’ on the melting column of a plume, a significant proportion of the melt will be derived from these more fusible, enriched components. In contrast, where the plume has a high flux of mass/buoyancy (e.g., Hawaii), or where the plume comes up at a ridge (e.g., Iceland), higher degrees of melting will result, and the enriched component in the melt becomes progressively diluted by melting of the more depleted, refractory component. The fact that picrites (which are generally regarded as representing a high degree of partial melting) are commonly the most depleted lavas in both the oceanic and continental realms provides support for this model. The predominance of depleted lavas in the Archean and more enriched lavas at many present-day hotspots led Campbell and Griffiths (1992) to suggest that the plume source changed from depleted to enriched throughout geologic time. Our model does not, however, require any significant change in the composition of the plume source region with time. Because the correlation between depleted and high-MgO (.18 wt%) lavas extends into the Archean, we suggest that the fundamental factor in determining the nature of the erupted melt from the plume is in fact the degree of partial melting. In the Archean the ambient mantle temperature was most likely hotter than it is now, and so plumes in the past would have undergone more extensive melting than they do today (Campbell and Griffiths, 1990). It is therefore possible that the mantle source region of Archean plumes was similar in composition to that of today, but 845

more extensive partial melting in the Archean led to formation of relatively abundant, high-MgO, depleted lavas. The point made by Anderson (1994) that ‘‘hotspots and MORB share a common depleted source’’ has no more meaning than saying that they both originate in the mantle. We propose that plumes can act as carriers of depleted material from the lower to the upper mantle; this depleted component may be isotopically distinct from the MORB reservoir, suggesting that it has had a distinct long-term history. Mixing into the upper mantle of the residue left after melting this ancient depleted plume component at hotspots may help to replenish the MORB source. The recognition of different depleted components has profound implications for geochemical mantle models. ACKNOWLEDGMENTS Supported by Natural Environment Research Council (NERC) Grant GR3/8984 to Tarney and Saunders and NERC studentships to Hards and Berry. We thank Bob Thompson for discussions, and Ray Kent, Robert Duncan, and an anonymous reviewer for helpful comments. REFERENCES CITED Anderson, D. L., 1994, Komatiites and picrites: Evidence that the ‘plume’ source is depleted: Earth and Planetary Science Letters, v. 128, p. 303–311. Budahn, J. R., and Schmitt, R. A., 1985, Petrogenetic modelling of Hawaiian tholeiitic basalts: A geochemical approach: Geochimica et Cosmochimica Acta, v. 49, p. 67– 87. Campbell, I. H., and Griffiths, R. W., 1990, Implications of mantle plume structure for the evolution of flood basalts: Earth and Planetary Science Letters, v. 99, p. 79–93. Campbell, I. H., and Griffiths, R. W., 1992, The changing nature of mantle hotspots through time: Implications for the chemical evolution of the mantle: Journal of Geology, v. 92, p. 497–523. Davies, G. F., and Richards, M. A., 1992, Mantle convection: Journal of Geology, v. 100, p. 151–206. Duncan, A. R., Armstrong, R. A., Erlank, A. J., Marsh, J. S., and Watkins, R. T., 1990, MORB-related dolerites associated with the final phases of Karoo flood basalt volcanism in southern Africa, in Parker, A., et al., eds., Mafic dykes and emplacement mechanisms: Rotterdam, Netherlands, Balkema, p. 119–129. Ellam, R. M., and Cox, K. G., 1991, An interpretation of Karoo picrite basalts in terms of interaction between asthenospheric magmas and the mantle lithosphere: Earth and Planetary Science Letters, v. 105, p. 330–342. Elliott, T. R., Hawkesworth, C. J., and Grønvold, K., 1991, Dynamic melting of the Iceland plume: Nature, v. 351, p. 201–206. Fitton, J. G., and James, D., 1986, Basic volcanism associated with intraplate linear features: Royal Society of London Philosophical Transactions, v. A317, p. 253–266. Griffiths, R. W., and Campbell, I. H., 1990, Stirring and structure in mantle starting plumes: Earth and Planetary Science Letters, v. 99, p. 66–78. Hards, V. L., Kempton, P. D., and Thompson, R. N., 1995, The heterogeneous Iceland plume: New insights from the alkalic basalts from the Snaefell volcanic centre: Geological Society of London Journal, v. 152 (in press). Hart, S. R., Hauri, E. H., Oschmann, L. A., and Whitehead, J. A., 1992, Mantle plumes and entrainment: Isotopic evidence: Science, v. 256, p. 517–520. Hauri, E. H., Whitehead, J. A., and Hart, S. R., 1994, Fluid dynamic and geochemical aspects of entrainment in mantle plumes: Journal of Geophysical Research, v. 99, p. 24,275–24,300. He´mond, C., Arndt, N. T., Lichtenstein, U., Hofmann, A. W., Oskarsson, N., and Steinthorsson, S., 1993, The heterogeneous Iceland plume; NdSr-O isotopes and trace element constraints: Journal of Geophysical Research, v. 98, p. 15,833–15,850. Hofmann, A. W., and White, W. M., 1982, Mantle plumes from ancient oceanic crust: Earth and Planetary Science Letters, v. 57, p. 421– 436. Holm, P. M., Gill, R. C., Pedersen, A. K., Larsen, J. G., Hald, N., Nielsen, T. F. D., and Thirlwall, M. F., 1993, The Tertiary picrites of West Greenland: Contributions from ‘Icelandic’ and other sources: Earth and Planetary Science Letters, v. 115, p. 227–244. Kerr, A. C., 1995, The melting processes and composition of the North Atlantic (Icelandic) plume; geochemical evidence from the early Tertiary lavas: Geological Society of London Journal, v. 152 (in press). 846

Kerr, A. C., Marriner, G. F., Arndt, N. T., Tarney, J., Nivia, A., Saunders, A. D., Storey, M. J., and Duncan, R. A., 1995, The petrogenesis of komatittes, picrites and basalts from the Isle of Gorgona, Colombia; new field, petrographic and geochemical constraints: Lithos (in press). Kesson, S. E., and Fitzgerald, J. D., 1991, Partitioning of MgO, FeO, NiO, MnO and Cr2O3 between magnesian silicate perovskite and magnesiowustite: Implications for the origin of inclusions in diamond and the composition of the lower mantle: Earth and Planetary Science Letters, v. 111, p. 229–240. Lawver, L. A., and Mu ¨ller, R. D., 1994, The Iceland hotspot track: Geology, v. 22, p. 311–314. Lightfoot, P. C., Hawkesworth, C. J., Hergt, J. M., Naldrett, A. J., Gorbachev, N. S., Fedorenko, V. A., and Doherty, W., 1993, Remobilisation of the continental lithosphere by a mantle plume: Major-, trace-element, and Sr-, Nd-, and Pb-isotopic evidence from picritic and tholeiitic lavas of the Noril’sk District, Siberian Trap, Russia: Contributions to Mineralogy and Petrology, v. 114, p. 171–188. MacDougall, J. D., 1988, Continental flood basalts and MORB; a brief discussion of similarities and differences in their petrogenesis, in MacDougall, J. D., ed., Continental flood basalts: Dordrecht, Netherlands, Kluwer, p. 331–341. Mahoney, J. J., 1988, Deccan Traps, in MacDougall, J. D., ed., Continental flood basalts: Dordrecht, Netherlands, Kluwer, p. 151–194. O’Nions, R. K., and Clarke, D. B., 1972, Comparative trace element geochemistry of Tertiary basalts from Baffin Bay: Earth and Planetary Science Letters, v. 15, p. 436– 446. Phipps-Morgan, J., and Shearer, P. M., 1993, Seismic constraints on mantle flow and topography of the 660-km discontinuity: Evidence for wholemantle convection: Nature, v. 365, p. 506–511. Richards, M. A., and Griffiths, R. W., 1989, Thermal entrainment by deflected plumes: Nature, v. 342, p. 900–902. Saunders, A. D., Norry, M. J., and Tarney, J., 1988, Origin of MORB and chemically-depleted mantle reservoirs: Trace element constraints, in Menzies, M. A., and Cox, K. G., eds., Oceanic and continental lithosphere: Similarities and differences: Journal of Petrology Special Lithosphere Issue, p. 415– 445. Schilling, J.-G., 1973, Icelandic mantle plume: Geochemical study of the Reykjanes Ridge: Nature, v. 242, p. 565–571. Schilling, J.-G., Meyer, P. S., and Kingsley, R. H., 1982, Evolution of the Iceland hotspot: Nature, v. 296, p. 313–320. Stille, P., Unruh, D. M., and Tatsumoto, M., 1986, Pb, Sr, Nd and Hf isotopic constraints on the origin of Hawaiian basalts and evidence for a unique mantle source: Geochimica et Cosmochimica Acta, v. 50, p. 2303–2319. Sun, S.-s., Tatsumoto, M., and Schilling, J.-G., 1975, Mantle plume mixing along the Reykjanes Ridge axis: Lead isotopic evidence: Science, v. 190, p. 13–17. Tackley, P. J., Stevenson, D. J., Glatzmaier, G. A., and Schubert, G., 1993, Effects of an endothermic phase transition at 670 km depth in a spherical model of convection in the Earth’s mantle: Nature, v. 361, p. 699–704. Thirlwall, M. F., Upton, B. G. J., and Jenkins, C., 1994, Interaction between continental lithosphere and the Iceland plume; Sr-Nd-Pb isotope geochemistry of Tertiary basalts, NE Greenland: Journal of Petrology, v. 35, p. 839– 879. Weaver, B. L., 1991, The origin of ocean island basalt end-member compositions: Trace element and isotopic constraints: Earth and Planetary Science Letters, v. 104, p. 381–397. White, R. S., 1988, A hot-spot model for the early Tertiary volcanism in the N Atlantic, in Morton, A. C., and Parson, L. M., eds., Early Tertiary volcanism and the opening of the NE Atlantic: Geological Society of London Special Publication 39, p. 3–13. White, R., and McKenzie, D., 1989, Magmatism at rift zones: The generation of volcanic continental margins and flood basalts: Journal of Geophysical Research, v. 94, p. 7,685–7,729. White, W. M., 1985, Sources of oceanic basalts and mantle evolution: Geology, v. 13, p. 115–118. Zindler, A., and Hart, S. R., 1986, Chemical geodynamics: Annual Review of Earth and Planetary Sciences, v. 14, p. 493–571. Zindler, A., Staudigel, H., and Batiza, R., 1984, Isotope and trace element geochemistry of young Pacific seamounts: Implications for the scale of upper mantle heterogeneity: Earth and Planetary Science Letters, v. 70, p. 175–195. Manuscript received March 6, 1995 Revised manuscript received June 8, 1995 Manuscript accepted June 21, 1995

Printed in U.S.A.

GEOLOGY, September 1995